Studia Informatica Universalis, Special Issue, Verification and Validation of multi-agent models for complex systems

website : http://lifc.univ-fcomte.fr/v2csissue2012/

Many scientists, from different areas of interest, exploit multi-agent systems (MAS) to model and/or to simulate complex systems such as crowd dynamic, social network and so on. Creating an agent based models conducts to make hypothesis because of the lack of knowledge about the studied system. In addition, a modelling process needs to establish models with different goals, based on various meta-models and languages. Thus model transcription problems and inconsistency between models may happen during this process.
Nevertheless, several researches present the model and simulators as idealistic tools needed for preventing disasters, for decision making, for understanding the functioning of complex systems or for imaging future evolutions of a complex system. But how confidence can we give to a model (and a simulator) ?
Due to the used modelling process, its internal complexity, an agent based model are rather complex to validate and/or to verify. Given from empirical comparison between simulation results and real data, many researches argue the efficiency of their models. Could we say that these models are valid ?
Several works investigate the domain of validation and verification of Agent Based model. They provide techniques to ensure the efficiency of model in general or specific case. There is two levels in this work :

The aim of this issue is to share experiences in validation and verification of complex system agent based model in order to make an overview of the domain. Every work and experience about retrospective validation, complex system case studies, validation and verification techniques are invited in this issue.